[openib-general] osm_state_mgr_process and osm_sm_state_mgr_process

[Hal Rosenstock]

On Tue, 2006-11-07 at 07:36, Eitan Zahavi wrote:
> Michael Arndt wrote:
> > Hi,
> >
> > is there an idea or concept why the state managment is seperate in these two 
> > functions.
> > I would really like if the answer isn't just yes or no.:)
> >   

This predates my involvement with OpenSM but I have the following
comments:

> The osm_state_mgr_process file is describing the OpenSM state machine.

This state machine handles the lower level aspects of the SM state
machine as detailed in IBA 1.2 vol 1 section 14.4. This contains
behavior "beyond the spec" in that much of SM behavior is an exercise
left to the reader.

> The osm_sm_state_mgr_process latter defines the SM state machine 
> following the IBTA SM states and their transitions.

This machine handles the high level aspects of the SM state machine in
14.4.1 and can be driven by the SMInfo attribute.

> The first one is far more complex and defines the stages required to 
> discover and configure the fabric.
> The second one is used to track the SM state in terms that are IBTA 
> compliant and standard. They are also map directly to SM states 
> available in the SMInfo attribute.

As to why they are separate state machines (and the implied question of
whether they could/should be combined), I'm not sure.

-- Hal

> > Thanks Michael
> >  
> >
> >
> > _______________________________________________
> > openib-general mailing list
> > openib-general at openib.org
> > http://openib.org/mailman/listinfo/openib-general
> >
> > To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
> >   
> 
> 
> _______________________________________________
> openib-general mailing list
> openib-general at openib.org
> http://openib.org/mailman/listinfo/openib-general
> 
> To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
>